>Mistral Code Enterprise is a fork of Continue. All due credit to the original creators of Continue.
Here's the problem with AI company revenue, I want this to be local, Continue is open source, my company laptop is powerful enough to run a usefully sized LLM locally more or less just as fast, and there are plenty of open models available publicly (including by Mistral, latest dev related called devstral).
As hardware progresses and LLM efficiency increases, I really don't think LLM programming has any future doing remote execution unless maybe as a chatbot kind of experience, but definitely not in an IDE.
Putting together a good developer experience with the open-ish and free-ish things available is still an exercise in cobbling a bunch of things together, but it's going to be streamlined soon and finding a way to compete with "zero cost and local" is going to be very hard for AI businesses.
Here's the problem with AI company revenue, I want this to be local, Continue is open source, my company laptop is powerful enough to run a usefully sized LLM locally more or less just as fast, and there are plenty of open models available publicly (including by Mistral, latest dev related called devstral).
As hardware progresses and LLM efficiency increases, I really don't think LLM programming has any future doing remote execution unless maybe as a chatbot kind of experience, but definitely not in an IDE.
Putting together a good developer experience with the open-ish and free-ish things available is still an exercise in cobbling a bunch of things together, but it's going to be streamlined soon and finding a way to compete with "zero cost and local" is going to be very hard for AI businesses.